A tool to load bulk simulation data into AWS IoT Analytics for experimenting with the capabilities of AWS IoT Analytics. This tool is intended to load simulation data that can then be utilized to explore the capabilities of Datasets, integration with Amazon QuickSight, etc.
Installation:
- Make sure you have a working Python environment
-
pip install onica-iotloaderto get the latest version installed
Template
A user-provided template is invoked to create each sample message. Arbitrary Python code can be used to generate each message via this template, enabling a rich set of simulated values. The template file must set a local variable named message to the value of a single simulated message, each time it is invoked. See template.py in this repository as an example of a static, but very large (~ 127kb) message template.
Usage
onica-iotloader [--concurrency=<n>] <channel> <template> <count>
A concurrency value of 1000 seems to be ideal to maximize the ingest rate of AWS IoT Analytics from a single source (achieving approximately 3200 msgs/sec, or 400 mb/sec on an m5.24xlarge). The <channel> must be an existing AWS IoT Analytics channel, <template> is the template Pyton script described above, and <count> is the number of messages (within a margin of error) to transmit. For reference, a count of 900000 results in just over 100GB of ingest using the default template.py included herein.
To load about 100GB in 4 minutes, run the following command using the template.py included in this repository (tested on m5.24xlarge):
onica-iotloader --concurrency=1000 <channel> template.py 900000
Replace <channel> with the name of an AWS IoT Analytics Channel configured in your account.
AWS Credentials & Region
The default boto3 AWS credential & region resolution (env vars, EC2 metadata, config file, etc) is utilized.
Feedback
Contact us at opensource at onica.com, or via Github issues, with feedback!
